Re: Problems with 0.3.2

[radford@robby.caltech.edu (Jim Radford)]

> 1 I have a IMAP group with gnus-info-read=((3 . 16)) which is correct
> 2 I start Gnus, it says that I have 2 new articles in that folder 
> 	(namely, 1 and 2)
> 3 If I select that folder it (correctly) says "No unread news" since
> 	article 1 and 2 doesn't exist
> 4 Gnus then believes it has read article 1 and 2, hence 
> 	gnus-info-read=((1 . 16))
> 5 nnimap-request-update-info updates this to ((3 . 16)) again
> 6 goto 1

> I have searched through all nnimap but I can't find any place that it
> tells Gnus that article 1 and 2 should exist. Any ideas?

When close-group is called with read ((1 . 16)) it tries to set the
\Seen mark for articles 1 and 2 which of course don't exist.  This is
OK except next time you compute the marks you don't get any marks for
non-existent articles, which for "read" means unread.  I guess that as
long as Gnus knows that the lowest article is 3, which it should if
you use the 'slow method for request list, then things should work.
So the question is: Does Gnus know this?  If it does then there is a
bug in Gnus.  If it doesn't (which is the case if you use any method
besides 'slow) then we can assume that articles below the lowest
article that exists are read, and mark them that way.  This should be
easy.

> >  o Gnus has other information besides read, ticked, and replied that
> >    would be usefull to keep on the server.  Any ideas on where to
> >    store it?

> IMAP support MUA-specific article marks, which we could use. Exactly
> what information where you thinking about?

Gnus has a big list of marks including: bookmark, dormant